The story
From a whiteboard idea to agents in production.
Two minutes of reading, thirty seconds of watching.
Turn ideas into AI agents that work.
Build, orchestrate, and scale AI agents on a visual canvas. Connect any model, your tools and your data, then run it on your own infrastructure.
Ideas → Agents → Automation → Real impact
Customer Support Agent
Automate support, analyze sentiment, escalate when needed.
Trigger
New ticket
Classify
with AI
Take Action
Reply / Escalate
Update
CRM
Waiting for a trigger…
0+
Nodes & integrations
0
Chat model providers
0
Vector stores
0
Data loaders
How ideas become impact
Four steps. Zero glue code.
01 · Ideas
It starts with an idea.
Answer every customer in seconds. Research a market overnight. Turn a backlog of documents into answers. Every team has ideas for AI, and most of them never leave the whiteboard.
02 · Agents
Shape it into agents.
Drag an LLM, give it tools, memory and your company knowledge. Connect agents to each other: a supervisor that delegates, a researcher that searches, a writer that drafts.
03 · Automation
Orchestrate the whole workflow.
Add conditions, loops, human approvals and API calls. Trigger from a webhook, a schedule or a chat. Watch every execution step by step and fix what needs fixing.
04 · Real impact
Ship it. Scale it. Own it.
Publish as an API, embed a chatbot on your site, or let your team use it in-app. It runs on your infrastructure, scales with workers and queues, and your data never leaves your control.
Features
Everything you need to put agents to work.
From the first prototype to thousands of executions a day.
Visual multi-agent orchestration
Design supervisors, workers and sub-agents on a canvas. Branch with conditions, loop over lists and hand off between agents without writing glue code.
300+ building blocks
LLMs, tools, memory, retrievers, document loaders and vector stores, ready to drag in.
Knowledge that stays fresh
Document stores ingest files, sites and databases, chunk and embed them, and keep your agents grounded in your data.
Human in the loop
Pause for approval before an agent sends, buys or deletes. Resume with feedback.
Executions you can see
Trace every run node by node: inputs, outputs, tool calls, tokens and timing.
API, SDK & embeddable chat
Every flow is an API. Embed a chatbot on any site or share a full-page assistant in one click.
Runs on your infrastructure
Docker, any cloud or on-prem. Postgres, S3 and Redis queue mode with workers when you need to scale.
Security built in
Encrypted credentials, API keys, rate limits, domain allow-lists and secret storage in AWS Secrets Manager.
Any model, your keys
OpenAI, Anthropic, Google, Azure, Bedrock, Mistral, Groq, Ollama and more. Swap models without rebuilding.
Use cases
Built for the work your team actually does.
Resolve tickets before anyone opens them
Classify incoming tickets, answer from your help center, issue routine refunds with approval, and escalate the hard ones with a ready summary.
- Sentiment-aware routing
- Answers grounded in your docs
- Human approval for refunds
Integrations
Bring your models, data and tools.
Swap providers in a click. No lock-in to any single AI vendor.
OpenAIAnthropicGoogle GeminiAzure OpenAIAWS BedrockMistralGroqOllamaCohereHugging FaceDeepSeekxAIOpenAIAnthropicGoogle GeminiAzure OpenAIAWS BedrockMistralGroqOllamaCohereHugging FaceDeepSeekxAI
PineconeQdrantPostgres pgvectorWeaviateChromaMilvusRedisMongoDB AtlasElasticsearchSupabaseZepUpstashPineconeQdrantPostgres pgvectorWeaviateChromaMilvusRedisMongoDB AtlasElasticsearchSupabaseZepUpstash
SlackGmailGoogle DriveGoogle SheetsMicrosoft TeamsOutlookNotionGitHubJiraConfluenceStripeMCP serversSlackGmailGoogle DriveGoogle SheetsMicrosoft TeamsOutlookNotionGitHubJiraConfluenceStripeMCP servers
Get started
Running in minutes, on your terms.
1
Create your account
Pick a plan and get your license key instantly. 14-day free trial, no card required.
2
Install anywhere
One Docker command on your cloud, data center or laptop. Postgres, S3 and Redis when you scale.
3
Build and ship agents
Start from templates or a blank canvas, connect your models and data, and publish as an API or chatbot.
# run AgentFactory with your license key docker run -d -p 3000:3000 \ -e AGENTFACTORY_LICENSE_KEY=<your-key> \ alexocloudai/agentfactory
Pricing
Simple plans. Self-hosted. Yours.
Every plan starts with a 14-day free trial.
Enterprise
For organizations with fewer than 1,000 employees.
$999/month
Start 14-day free trial- Unlimited self-hosted instances
- Unlimited users
- Visual agent & workflow builder
- Queue mode for scale (workers + Redis)
- Executions & observability
- Priority support
Most popular
Enterprise Plus
For organizations with up to 3,000 employees.
$2,999/month
Start 14-day free trial- Unlimited self-hosted instances
- Unlimited users
- Visual agent & workflow builder
- Queue mode for scale (workers + Redis)
- Executions & observability
- Priority support
- Deployment assistance
Enterprise Custom
For organizations with more than 3,000 employees.
Custom
Talk to sales- Unlimited self-hosted instances
- Unlimited users
- Visual agent & workflow builder
- Queue mode for scale (workers + Redis)
- Executions & observability
- Priority support
- Offline / air-gapped licensing
- Dedicated success manager
- SLA-backed support
Need something different? Talk to sales.
FAQ
Questions, answered.
Yes. You download AgentFactory and run it on your own servers or cloud account with Docker. Your flows, credentials and data stay in your infrastructure.
Turn ideas into AI agents that work.
Start a 14-day free trial. Install in minutes, keep your data on your own servers.